Null Pointer Error while doing “alter table table_name drop partition(part_column < value)” on Hive CLI

Question

I have a spark job (Scala) which writes time-series data onto Hadoop over which there is an external table in Hive.

The table is partitioned by multiple columns and one of the columns ( circle ) has spaces in its values(eg "Punjab and Rajasthan" ).

Within the spark job when I try to do

sparksession.sql("""
alter table table_name 
drop if exists partition(creation_time < latestcreationtime)
"""
)

I get an Illegal Character Exception of hive metastore, stack trace is attached. I get the same in hive CLI, looks as Hive is unable to read spaces.

User class threw exception: org.apache.spark.sql.AnalysisException: org.apache.hadoop.hive.ql.metadata.HiveException: MetaException(message:Illegal character in path at index 131: /user/optimus/rohit/hive_dump/c360/version=v1.28/set_name=d_si/creation_time=1610994976/compaction_flag=U/si_lob=DTH/circle=Andaman and Nicobar Islands);

Answer 1

1.Try dropping and recreating the table 2.Your data might have some issue, what is the underlying data type and table type?